Xiao-peng Wang,Jing Li,Yue Liu.Watershed segmentation based on gradient relief modification using variant structuring element[J].Optoelectronics Letters,2014,10(2):152-156
Watershed segmentation based on gradient relief modification using variant structuring element
Author NameAffiliationE-mail
Xiao-peng Wang School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou, 730070, China wangxp1969@sina.com 
Jing Li School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou, 730070, China  
Yue Liu School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou, 730070, China  
Abstract:
      Watershed segmentation is suitable for producing closed region contour and providing an accurate localization of object boundary. However, it is usually prone to over-segmentation due to the noise and irregular details within the image. For the purpose of reducing over-segmentation while preserving the location of object contours, the watershed segmentation based on morphological gradient relief modification using variant structuring element (SE) is proposed. Firstly, morphological gradient relief is decomposed into multi-level according to the gradient values. Secondly, morphological closing action using variant SE is employed to each level image, where the low gradient level sets use the large SE, while the high gradient level sets use the small one. Finally, the modified gradient image is recomposed by the superposition of the closed level sets, and watershed transform to the modified gradient image is done to implement the final segmentation. Experimental results show that this method can effectively reduce the over-segmentation and preserve the location of the object contours.
Hits: 3917
Download times: 0
This work has been supported by the National Natural Science Foundation of China (No.61261029), the Jinchuan Company Research Foundation (No.JCYY2013009), the Gansu Province Science and Technology Support Program (No.1204GKCA051), and the Gansu Province Higher Education Fundamental Research Funds (No.212090).
View Full Text    Download reader